In order that the school test of mental deficiency should be as good as the Binet system it would have to provide a standard of school progress relative to length of attendance instead of school position relative to age. If one could say that a child was not above the lowest 0.5% of the children of his age in the _progress_ which he had made in school relative to the time actually spent in school, one would then have an excellent standard for judging feeble-mindedness for any child who had been in school for some years. It would be better if an uncertain region were also defined. By the time that a child's ability has been passed upon for four or five years and by different teachers, even from the point of view of the needs of school work, one has a criterion for mental ability in a particular community applied under long observation, which no system of brief tests can hope to equal for some time to come. Such a standard, however, is unfortunately not available since we have too little information about school progress relative to attendance. Even if it were available, psychological tests would still be an important check upon the school judgment on account of the excessive value put upon mere memorizing in school and on account of the emotional repulsion to the school developed by some children of ability. Mental tests would be necessary, moreover, for the younger ages. (b) SCHOOL RETARDATION AS A WARNING OF THE NEED FOR EXAMINATION. Even if no more is known than a person's grade in school at any age over eleven it is an important cue as to his mentality. Here our problem is not estimating deficiency among groups but the discovery of deficient individuals. We wish to find the highest grade in school in which we are at all likely to find children under present conditions who test in the lowest 1.5% for their ages. Our records on 653 15-year-olds indicate that a pupil of this age who tests doubtful is very rarely retarded less than 3 years in school. It occurred only twice when tested ability was judged by the 1911 tests, four times judged by the 1908 scale. None of the 15-year-olds who tested presumably deficient were retarded less than three years. In Minneapolis, as in many cities, the custom prevails of promoting, regardless of passable work, after two years have been spent in a grade. We suggest, therefore, to be perfectly safe, it is well for every child in court to be examined who is two years retarded in school below the standard age of 7 in the first grade and is not able to carry work above the seventh grade. This will include a considerable number of children at the lower border of those presumably passable.